On Oct 28, 2005, at 5:35 PM, Troy Telford wrote:
Excellent. Many thanks for all your diligence here -- this is
extremely helpful to us!
Believe me, I have a pretty big interest in helping OpenMPI succeed.
Much
of what OpenMPI can already do solves so many things.
Glad to hear it; this is one of the reasons Open MPI exists.
When we all first started talking about collaborating (just about 2
years ago), we initially thought we'd just write some common components
for each of our MPI implementations (LAM/MPI, LA-MPI, FT-MPI). But we
eventually came to the realization that each of our projects had good
*and* bad points, and it was strategically a much better move -- both
for us and for the greater HPC community -- to just completely start
over with a blank slate and as few preconceptions as possible. Hence,
Open MPI was born.
The point here is that if/when you find things that Open MPI
does/doesn't do that is a point point for you, please be sure to let us
know.
Can you grab either the latest v1.0 SVN
or the v1.0 snapshot and give it a whirl?
Sure, I'll give it a whirl.
Thanks!
Just out of curiosity -- do you test OpenMPI for memory leaks using
Valgrind (or similar)?
Yes. Unfortunately, we've still got some work to do there, too. :-\
--
{+} Jeff Squyres
{+} The Open MPI Project
{+} http://www.open-mpi.org/